Beefy Boxes and Bandwidth Generously Provided by pair Networks
good chemistry is complicated,
and a little bit messy -LW
 
PerlMonks  

Re^2: Where to put self-made loop logic (separate module)?

by basic6 (Novice)
on Dec 30, 2012 at 16:07 UTC ( #1010936=note: print w/ replies, xml ) Need Help??


in reply to Re: Where to put self-made loop logic (separate module)?
in thread Where to put self-made loop logic (separate module)?

Wow, so many answers already. :-)

The best way to find out if this should be a separate module is probably to let you guys decide, so here's the code...
This code parses some special list markup (similar to Wiki markup).
The scalarref ($text) is directly modified (I don't want to copy it and go $$text = $newtext).
If I wanted to write another parser function which does even more complicated modifications (which cannot be replaced by a regex), I'd have some duplicate code if this loop was not a separate module.
I know this isn't the prettiest piece of code mankind has written, but it seems to work (even with 3 byte UTF-8 characters / note that the line length changes).
I'd be happy if you could help me improve my code (even if there's a one-line alternative).

sub parseLists { my $self = shift; my $text = \$self->{_text}; # So we want to loop through the text line by line # and be able to modify some lines, # but we don't want to rebuild/copy the whole text. my $lf = "\n"; # linebreak my $lflen = length $lf; # 1 my $pos1 = 0; # left line offset my $pos2 = 0; # right line offset (1st char after line) my $len = 0; # line length my $lendif = 0; # line length difference my $inlist; open my $fh, "<:utf8", $text; # Note how we open in UTF-8 mode # while (<$fh>) # Gets confused when line length changes # Using seek() is risky, because it reads bytes, not chars! # However, substr() always counts chars, not bytes. while (<$fh>) { # Get line string without newline character my $line = substr $_, 0, -$lflen; my $oldline = $line; # Calculate offsets $len = length $line; $pos1 = $pos2; $pos2 += $len + $lflen; # Modify line # START (not part of loop structure) my $isasterisk = $line =~ m/^\* /; my $isindented = $line =~ m/^\ /; my $isfirst; if (!$inlist) { if ($isasterisk) { $isfirst = 1; $inlist = 1; } } if ($inlist) { if (!$isindented && !$isasterisk) { substr $line, 0, 0, "</ul>\n"; $inlist = 0; } elsif ($isindented) { $line =~ s/^\ (.*)/<li class="nobullet">$1<\/li>/; } elsif ($isasterisk) { $line =~ s/^\* (.*)/<li>$1<\/li>/; substr $line, 0, 0, "<ul>\n" if $isfirst; } } # END # Write new line back substr $$text, $pos1, $len, $line; # Calculate diff $lendif = (length $line) - ((length $_) - ($lflen)); # Adjust our and Perl's (!) position counter $pos2 += $lendif; # That's our counter seek $fh, $lendif, 1; # That's from Perl / SEEK_CUR } }


Comment on Re^2: Where to put self-made loop logic (separate module)?
Download Code
Re^3: Where to put self-made loop logic (separate module)?
by davido (Archbishop) on Dec 30, 2012 at 17:56 UTC

    Does the format of the input data have a name? Is the input data format documented somewhere?


    Dave

      The input is (part of) a template. In other words, it's just human-readable text, no code (no Perl, no HTML). There may be a couple of lines, each starting with an asterisk - that's what my function parses.
      But that's probably not important, as my goal is to have a line-by-line loop structure which allows in-place modifications. For that matter, the input is guaranteed to always be UTF-8 encoded text (only printable characters and whitespaces).

        With the parsing logic embedded, it seems like a very specific tool that targets an uncommon input format. That's not "generally useful", and probably doesn't make for a great CPAN candidate.

        If you write the loop such that one could pass in a subref, it might be a more generalized tool, but then it's just a loop, and I don't see much advantage to it.

        This isn't to say you're not onto something. I just think it would require that you decide what exactly you want your module to accomplish. Throughout this thread I just see a module that handles a specific type of data that nobody else is likely to need to handle.


        Dave

Log In?
Username:
Password:

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://1010936]
help
Chatterbox?
and the web crawler heard nothing...

How do I use this? | Other CB clients
Other Users?
Others lurking in the Monastery: (9)
As of 2014-11-24 21:59 GMT
Sections?
Information?
Find Nodes?
Leftovers?
    Voting Booth?

    My preferred Perl binaries come from:














    Results (147 votes), past polls